Fig. 7

Relationship between inverted repeat (IR) length and distance between repeats. (a) is from 50 validated NAHR inversions from invFest and (b) is from 30 predicted NAHR inversions at NA12878 using npInv. Two points are drawn for each pair of IR which mediates a NAHR inversion. In the majority of cases the pair of IRs have similar length and so the two points are co-localised; otherwise the two points are connected by a line. The distance between repeats in NAHR inversion is the inversions size without repeats. The y-axis indicates the distance between the IRs and the x-axis indicates the length of the IR. The dotted line indicates the linear regression from log10 average size of IR to log10 size of inversion without flanking IR. The green long dash lines are 90% prediction intervals. The slope, intercept, R-squared and t-test p-value for linear regression are shown at bottom right. The background heat map indicates the count of IR in the genome with given IR length and distance between IR. The size of short interspersed nuclear elements (SINEs, 300 bp) and long interspersed nuclear elements(LINEs, 7 kb) is shown in the x-axis legend